- perform full sensor signal path reset and wait for max time (100 ms)
- issue full sensor reset on any error
- always read FIFO count before transfersj
- only allocate drdy perf counter if GPIO is available
- IMU integration move from drivers (PX4Accelerometer/PX4Gyroscope) to sensors/vehicle_imu
- sensors: voted_sensors_update now consumes vehicle_imu
- delete sensor_accel_integrated, sensor_gyro_integrated
- merge sensor_accel_status/sensor_gyro_status into vehicle_imu_status
- sensors status output minor improvements (ordering, whitespace, show selected sensor device id and instance)
- bring in PX4/ecl#795 "EKF: Improve covariance prediction stability"
- the ecl/EKF filter update period has changed from 8 ms to 10 ms
- change default integration period 4000 us -> 2500 us (aligns with new EKF filter update period)
The InvenSense icm20608g and icm20689 don't have a FIFO watermark interrupt, but they do have a data ready interrupt and the ability to get the current FIFO count in the same large transfer as the actual FIFO data. So instead of manually checking the FIFO count before every transfer (costs ~ 1-3% cpu) we trust the data ready counts, verify things are in sync after the large transfer (fifo count + fifo contents), and force a manual check before the next transfer if necessary.
As a precaution this change lowers the threshold for forcing a manual FIFO count check before the large transfer. Forcing the check if out of sync by 2 (or more) samples makes sense because we always do transfers in multiples of 2 (gyro samples per accel).
The other small cosmetic changes are keeping the icm20608g and icm20689 in sync (they're nearly identical).
- mag enabled with -M when starting
- always check FIFO count before transfer
- interupt pin set active low and latch
- relax retry timeout if configure failed
- improve configured empty rate (sample rate) rounding
- fix RegisterCheck
- check FIFO count as part of full transfer and reset or adjust timing if necessary
- rename DRV_MAG_DEVTYPE_MPU9250 -> DRV_MAG_DEVTYPE_AK8963
- always check FIFO count before transfer even with data ready
- interupt pin set active low and latch
- relax retry timeout if configure failed
- improve configured empty rate (sample rate) rounding
- check FIFO count as part of full transfer and reset or adjust timing if necessary
- change default FIFO empty rate to 1 kHz (consistent with other drivers)
- relax retry timeout if configure failed
- improve configured empty rate (sample rate) rounding
- check interupt status and FIFO count as part of full transfer and reset or adjust timing if necessary
- change default FIFO empty rate to 1 kHz (consistent with other drivers)
- relax retry timeout if configure failed
- improve configured empty rate (sample rate) rounding
- check interupt status and FIFO count as part of full transfer and reset or adjust timing if necessary
- interupt pin set active low and latch
- relax retry timeout if configure failed
- improve configured empty rate (sample rate) rounding
- fix RegisterCheck
- check FIFO count as part of full transfer and reset or adjust timing if necessary
- interupt pin set active low and latch
- relax retry timeout if configure failed
- improve configured empty rate (sample rate) rounding
- fix RegisterCheck
- check FIFO count as part of full transfer and reset or adjust timing if necessary
- rename DRV_IMU_DEVTYPE_ICM20608 -> DRV_IMU_DEVTYPE_ICM20608G
- update to NuttX with stm32f4 and stm32f7 SPI DMA internal buffers
- remove explicit DMA buffer allocations from new IMU drivers
- restore original BOARD_DMA_ALLOC_POOL_SIZE
- decrease SPI DMA thresholds based on fmu-v2/v3/v4/v5 bench testing
This is another new InvenSense sensor with max output data rate of 32 kHz (both accel & gyro), on board anti-aliasing filter, and accel range 32G. Overall quite similar to the ICM-42688-P added in #14420, but with doubled accel range and without the optional high resolution output or clock sync.
This is a new InvenSense sensor with max output data rate of 32 kHz (both accel & gyro), on board anti-aliasing filter, optional higher resolution output (18 bit accel, 19 bit gyro), and clock sync.
This avoids the need for recalibration, and also cleans up other driver
ID's (merge separate accel/gyro).
The SPI address was previously set to a board-specific (arbitrary) value,
and is now set to 0. This will allow extending for multiple sensors of the
same type on the same bus.
- refactor Run() into simple state machine
- perform reset and configuration in sensor bus thread
- when using data ready interrupt skip checking FIFO count
- fix periodic temperature sampling (rate limit to 1 Hz)
- refactor Run() into simple state machine
- perform reset and configuration in sensor bus thread
- when using data ready interrupt skip checking FIFO count
- fix periodic temperature sampling (rate limit to 1 Hz)
- refactor Run() into simple state machine
- perform reset and configuration in sensor bus thread
- when using data ready interrupt skip checking FIFO count
- fix periodic temperature sampling (rate limit to 1 Hz)
- refactor Run() into simple state machine
- perform reset and configuration in sensor bus thread
- when using data ready interrupt skip checking FIFO count
- checked register mechanism and simple watchdog
- driver checks for errors gradually and can reconfigure itself
- respect IMU_GYRO_RATEMAX at the driver level
- fixed sensor INT16_MIN and INT16_MAX handling (y & z axis are flipped before publishing)
- checked register mechanism and simple watchdog
- driver checks for errors gradually and can reconfigure itself
- respect IMU_GYRO_RATEMAX at the driver level
- fixed sensor INT16_MIN and INT16_MAX handling (y & z axis are flipped before publishing)
- checked register mechanism and simple watchdog
- driver checks for errors gradually and can reconfigure itself
- respect IMU_GYRO_RATEMAX at the driver level
- fixed sensor INT16_MIN and INT16_MAX handling (y & z axis are flipped before publishing)
- checked register mechanism and simple watchdog
- driver checks for errors gradually and can reconfigure itself
- respect IMU_GYRO_RATEMAX at the driver level
- fixed sensor INT16_MIN and INT16_MAX handling (y & z axis are flipped before publishing)
- increased sensor_gyro_fifo max size (enables running the driver much slower, but still transferring all raw data)
- PX4Accelerometer/PX4Gyroscope remove unnecessary memsets
I've added a queue depth of 4 for sensor_accel and sensor_gyro. This is initially added because it's not always possible for the `vehicle_acceleration` to keep up with every publication of the primary accelerator as it runs in the same thread as ekf2, various controllers, etc.
Later this mechanism will be used in a few areas
- rate limit `vehicle_angular_velocity` and `vehicle_acceleration` without missing any raw data
- move IMU integration to `vehicle_imu` and out of the actual driver threads, eliminating the need for sensor_accel_integrated and sensor_gyro_integrated
- integrate raw gyro synchronized with optical flow measurements
- gyro filtering (low-pass and notch) only performed on primary gyro in `sensors/vehicle_angular_velocity` instead of every gyro in `PX4Gyroscope`
- sample rate is calculated from actual updates (the fixed value was slightly wrong in many cases, and very wrong in a few)
- In the FIFO case the array is now averaged and published in `sensor_gyro` for filtering downstream. I'll update this in the future to use the full FIFO array (if available), but right now it should be fine.
- this provides some extra space when the FIFO transfers don't align perfectly
- I've also made an effort to keep the different drivers (icm20602, icm20608g, ism330ldc) in sync so we can factor out the common portions later once we've confident in the pattern.
- uses the FIFO and SPI DMA to transfer full raw data (8 kHz gyro, 4 kHz accel)
- new sensor messages for better visibility
- sensor_{accel, gyro}_fifo: full raw data for optional logging and analysis
- sensor_{accel, gyro}_status: metadata, clipping, etc
- currently not enabled by default
Daniel Agar